Introduction

Congenital stationary night blindness (CSNB) is a group of clinically and genetically heterogeneous non-progressive retinal diseases. Nougaret-type CSNB with autosomal dominant inheritance is well described in a single large French family in which all affected patients descended from the first reported patient, Jean Nougaret [1, 2]. A heterozygous variant (p.G38D) in the GNAT1 gene, encoding a rod-specific transducin α-subunit, has been identified in patients with Nougaret-type CSNB [2]. At present, the family is the only known Nougaret-type CSNB family with GNAT1 variant (p.G38D). Other variants in three genes, rhodopsin (RHO) [3], rod-specific cGMP phosphodiesterase type 6 β-subunit (PDE6B) [4, 5], and GNAT1 [6,7,8], have been associated with “Riggs-type” autosomal dominant CSNB. Functional defects of these three genes affect the rod phototransduction cascade.

A large number of variants in the ABCA4 gene, encoding an ATP-binding cassette superfamily transmembrane protein, have been reported in patients with autosomal recessive Stargardt macular dystrophy (STGD1, MIM #248200) since its first description in 1997 [9]. Biallelic variants in the ABCA4 gene are responsible for a spectrum of retinal dystrophies (ABCA4-associated retinopathies) including STGD1, cone (COD), and cone-rod dystrophies (CORD) [10,11,12].

More than 270 other genes causing inherited retinal diseases (IRDs) have been cataloged in the Retinal Information Network (RetNet, https://sph.uth.edu/retnet/home.htm), and this database has grown in recent years with the application of whole exome sequencing (WES), a popular next-generation sequencing technology. This has allowed ophthalmologists to identify disease-causing gene variants not only in IRDs but also in undiagnosed retinal diseases.

To date, autosomal dominant CSNB with any known disease-causing gene variant has never been reported in the Japanese population. Recently, we encountered five patients with autosomal dominant CSNB and two patients with childhood-onset CORD in a Japanese family. The purpose of this study was to characterize clinical and genetic features of those patients. Genetic analysis using WES revealed that the CSNB patients carried the known heterozygous GNAT1 variant (p.G38D), while the CORD patients carried biallelic ABCA4 variants as well as the GNAT1 variant.

Methods

Ethical approval

The Institutional Review Boards of the Jikei University School of Medicine (approval no. 24-232 6997) and Hamamatsu University School of Medicine (approval no. 14-040) approved the protocol for this study. The protocol adhered to the tenets of the Declaration of Helsinki, with informed consent obtained from each participant or their legal guardians.

Clinical examinations

A total of nine family members from three different generations of a Japanese family were recruited in this study (Fig. 1A). Four (II-2, II-3, III-1, and III-2) of the nine were ophthalmologically assessed at the Jikei University Hospital. We performed a comprehensive ophthalmic examination, including decimal best corrected visual acuity (BCVA), slit-lamp examination, funduscopy, fundus autofluorescence imaging (FAF) using a Spectralis HRA (Heidelberg Engineering, Heidelberg, Germany), and an Optos 200Tx, Ultra-Wide Field Retinal Imaging System (Optos, Dunfermline, UK), and optical coherence tomography (OCT; Carl Zeiss Meditec AG, Dublin, CA, USA). Horizontal B-scan images through the fovea were obtained. Full-field electroretinography (FF-ERG) using a light-emitting diode built-in electrode (LE-4000, Tomey, Nagoya, Japan) was recorded in accordance with the protocols of the International Society for Clinical Electrophysiology of Vision [13]. Details on the procedure and experimental conditions have been previously reported [14]. ERG recording of the proband (II-2, JU#1542) was dark-adapted for 30 min or 24 h in either eye. On–off ERG with 100-ms stimulus duration was also recorded for the proband.

Fig. 1
figure 1

Pedigree of the Japanese family and nucleotide sequences of exon 2 of the GNAT1 gene. A Unaffected members are shown as unfilled circles (female) or squares (males), whereas patients with congenital stationary night blindness (CSNB) or cone-rod dystrophy (CORD) are shown as gray or black symbols, respectively. The female proband (II-2) is indicated by an arrow. The “wt” denotes wild-type genotype. A heterozygous GNAT1 variant (p.G38D) is found in all examined CSNB patients including the proband, whose nucleotide sequencing is shown in B, while the CORD patients (III-1) carry compound heterozygous ABCA4 variants (p.Q185X and c.1760 + 2 > G), whose nucleotide sequencing is shown in C, as well as the GNAT1 variant

Full size image

Whole exome sequencing

Genomic DNA was extracted from peripheral blood leukocytes using a Gentra Puregene Blood Kit (Qiagen, Hilden, Germany). WES was performed on four family members (II-1, II-2, III-1, and III-2; Fig. 1A). Our detailed WES methodology has been described elsewhere [14, 15]. To find out disease-causing variants, we focused on nonsynonymous variants and splice site variants, which are within 10 bp of exon–intron boundaries, and excluded synonymous and noncoding exonic variants from the analysis. We treated common genetic variants (allele frequency > 0.01 for recessive variants or > 0.001 for dominant variants) in any of the ethnic subgroups found in the following single-nucleotide polymorphism databases and synonymous variants as putative nonpathogenic sequence alterations: 1000 Genomes database (http://www.1000genomes.org/), Exome Aggregation Consortium database (http://exac.broadinstitute.org/), Genome Aggregation Database (https://gnomad.broadinstitute.org/), Human Genetic Variation Database (http://www.genome.med.kyoto-u.ac.jp/SnpDB/), and Tohoku Medical Megabank Organization database (https://ijgvd.megabank.tohoku.ac.jp/).

Sanger sequencing for validation and co-segregation analysis

Sanger direct sequencing was performed to confirm the detected variants and to conduct co-segregation analysis on eight family members (I-1, I-2, II-1, II-2, II-3, II-4, III-1, and III-2). Considering clinical features of patients, mode of inheritance in the family, and co-segregation analysis, disease-causing variants were finally confirmed from the detected variants in the IRD genes. For variants of the GNAT1 and ABCA4 genes, the following primer sets were used. GNAT1 exon 2: forward 5′-GGTCCACCCTGCCAACTC-3′, reverse 5′-ATAAACTCGAGGCACTCTTCCAG-3′; ABCA4 exon 5: forward 5′-TTACAAGTGTTTCCAATCGACTCT-3′, reverse 5′-GTTGAAAATGATCACGATCTCAAG-3′; ABCA4 exon 12: forward 5′-CTGGACACGTTGAAAAATTAACAC-3′, reverse 5′-ATTCAAGGACTCTTTGTTGAGCTT-3′. NCBI Reference Sequences of GNAT1 (NM_144499.2) and ABCA4 (NM_000350.2) were used.

Results

Congenital stationary night blindness (CSNB) phenotypes

Patient II-2

A 34-year-old female proband (II-2, JU#1542) was referred to The Jikei University Hospital for assessment of her congenital night blindness. She had never complained of visual symptoms except for night blindness, nor had systemic medical history. The condition of night blindness was stationary, and therefore, she was diagnosed with CSNB. Her decimal BCVA was 1.2 (− 1.75 diopter sphere) in the right and 1.5 (− 1.75 diopter sphere) in the left eye. Slit-lamp examination showed no abnormal findings in the anterior segment and media. No remarkable findings were observed in funduscopy, fundus autofluorescence (FAF), or cross-sectional macular OCT images (Fig. 2). FF-ERG (Fig. 3) showed non-recordable rod response to a weak flash (dark-adapted 0.01 cd s m−2; DA 0.01) in rod ERG but at least half of a- and b-wave responses in DA 3.0, DA 10.0, and DA 200 ERG after 30 min of DA in the right eye, as well as after 24 h of DA in the left eye. Notably, the DA 3.0 responses showed distinguishing delayed and broadened a-waves and negative-type waveforms (b/a ratio less than 1.0). The photopic cone (light-adapted 3.0 cd s m−2; LA 3.0) and 30-Hz flicker (LA 3.0 flicker) responses were completely normal. Clear on and off responses were observed.

Fig. 2
figure 2

Fundus photograph, fundus autofluorescence (FAF), and B-scan optical coherence tomography (OCT). No remarkable findings were present in fundus photographs (A), FAF (B), or OCT (C) of patients (II-2, II-3) with congenital stationary night blindness

Full size image
Fig. 3
figure 3

Full-field electroretinography (ERG). ERG findings of a control, patients (II-2 and II-3) with congenital stationary night blindness (CSNB), and patients (III-1 and III-2) with cone-rod dystrophy (CORD). ERG of CSNB patient II-2 shows non-recordable rod response (dark-adapted 0.01 cd s m−2; DA 0.01) but at least 50% greater than normal a- and b-wave responses in DA 3.0, DA 10.0, and DA 200 ERG after 30 min of DA in the right eye, as well as after 24 h of DA in the left eye. The cone (light-adapted 3.0 cd s m−2; LA 3.0) and 30-Hz flicker (LA 3.0 flicker) responses are completely normal. Clear on and off responses are observed. The ERG findings of CSNB patient II-3 are compatible with those of patient II-2. In CORD patients III-2 and III-1, DA ERG shows non-recordable rod response (DA 0.01), and severely decreased a- and b-wave responses to a strong flash (DA 3.0), but notably about one-third of normal response to a stronger flash (DA 200). The LA 3.0 and LA 3.0 flicker responses are severely decreased in III-2, consistent with features of CORD

Full size image

Patient II-3

A 37-year-old female patient (II-3, JU#1608, elder sister of the proband) was ophthalmologically evaluated and diagnosed with CSNB. Her decimal BCVA was 1.5 (− 1.00 diopter sphere) in the right and 1.5 (− 0.50 diopter sphere) in the left eye. No remarkable findings were observed in slit-lamp examination, fuduscopy, fundus autofluorescence (FAF), or cross-sectional macular OCT images (Fig. 2). The FF-ERG findings (Fig. 3) were essentially very similar to those of the proband.

Patients I-2, II-4, and II-5

Patients I-2 (JU#1594, a 74-year-old female, maternal aunt of the proband), II-4 (JU#1619, a 49-year-old female, maternal cousin of the proband), and II-5 (45-year-old female, maternal cousin of the proband), who had good visual acuities, had never complained of visual symptoms except for congenital night blindness. The diagnosis of these three patients was also compatible with CSNB. Incidence of the “CSNB” disease in this family was consistent with autosomal dominant inheritance (Fig. 1A).

Cone-rod dystrophy (CORD) phenotypes

Patient III-1

A 7-year-old female patient (III-1, JU#1540, daughter of the proband), who complained of congenital night blindness, was evaluated because of progressive deterioration of visual acuity. Her decimal BCVA was 0.35 (+ 2.50 diopter sphere) in the right and 0.2 (+ 3.25 diopter sphere) in the left eye, although her BCVA was nearly 1.0 at 5 years of age. Slit-lamp examination showed no remarkable findings. Retinal examination revealed gray discoloration within the vascular arcade in the fundus photographs, hypoautofluorescent macular area with a hyperautofluorescent ring in FAF, and disrupted ellipsoid zone (EZ) with foveal thinning in OCT (Fig. 4). Dark-adapted FF-ERG of the right eye (Fig. 3) showed non-recordable rod response to a weak flash (DA 0.01), severely decreased a- and b-wave responses to a strong flash (DA 3.0), but notably about one-third of normal response to a stronger flash (DA 200).

Fig. 4
figure 4

Fundus photograph, fundus autofluorescence (FAF), and B-scan optical coherence tomography (OCT). A Fundus photographs show gray discoloration within the vascular arcade in III-1 and no remarkable finding in III-2. B FAF shows hypoautofluorescent macular area with hyperautofluorescent ring in III-1 and hyperautofluorescent ring around the macula in III-2. C OCT shows disrupted ellipsoid zone (EZ) with foveal thinning in III-1. In III-2, OCT reveals a prominent hyperreflective line, corresponding to thickened external limiting membrane, underneath the outer nuclear layer at 5 years of age (5 years), but later disappearance of the line and thinner foveal thickness at 6 years, 8 months of age (6 years)

Full size image

Patient III-2

A 5-year-old male patient (III-2, #JU1566, son of the proband), who complained of congenital night blindness, was evaluated. At 5 years of age, his BCVA was 0.7 (+ 0.25 diopter sphere) in the right and 0.7 (+ 0.25 diopter sphere) in the left eye. Slit-lamp and funduscopy examination showed no remarkable findings (Fig. 4). FAF showed a hyperautofluorescent ring around the macula (Fig. 4). OCT showed a thickened external limiting membrane (ELM) as previously reported [16]. Later, when he was 6 years, 8 months of age, his BCVA decreased to 0.1 in the right and 0.15 in the left eye. Surprisingly, OCT revealed disappearance of the thickened ELM and thinner foveal thickness compared to that at 5 years of age (Fig. 4), demonstrating progressive loss of visual acuity. Dark-adapted FF-ERG findings (Fig. 3) were essentially very similar to those of patient III-1 (JU#1540). The photopic cone (LA 3.0) and 30-Hz flicker (LA 3.0 flicker) responses were severely decreased (Fig. 3), consistent with known features of CORD.

Identification of pathogenic variants of the GNAT1 and ABCA4 genes

We examined all known IRD genes to screen for variants from WES data, which we then analyzed based on the autosomal dominant, autosomal recessive, and de novo models. After synonymous variants were excluded, a total of nine rare variants remained (data not shown). We then selected remaining variants that matched the patients’ phenotypes and the disease phenotypes known to be caused by the IRD genes listed in the RetNet database. As a result of WES analysis, a heterozygous variant (c.113G > A, p.G38D) in the GNAT1 gene, in agreement with autosomal dominant inheritance, was identified in the proband (II-2) (Fig. 1B). The variant was also found in the CORD patients (III-1 and III-2), but not in the unaffected member (II-1). In addition, compound heterozygous ABCA4 variants, c.553C > T, p.Q185*, and c.1760 + 2T > G, were also identified in the CORD patients (Fig. 1A, C), whose parents (II-1 and II-2) had either of the variants (Fig. 1A), compatible with autosomal recessive inheritance. Sanger sequencing also confirmed these GNAT1 and ABCA4 variants. Co-segregation analysis revealed that the GNAT1 variant was found in all CSNB patients except for patient II-5 who did not undergo genetic testing, but not in the unaffected members (I-1, II-1) (Fig. 1A). The GNAT1 variant (p.G38D) has been reported only in Nougaret-type CSNB [2]. On the other hand, each of the nonsense (p.Q185*) and splice site (c.1760 + 2T > G) variants in ABCA4 has been reported in ABCA4-associated retinopathies [17, 18].

Discussion

In this study, we identified a previously documented heterozygous GNAT1 variant (p.G38D) in CSNB patients of a Japanese family. Also, we found coexistence of the GNAT1 and biallelic ABCA4 variants in two patients with both CSNB and childhood-onset CORD from this family.

To date, the GNAT1 variant has been found only in a large French family with Nougaret-type CSNB [2, 19]. Details of characteristic ERG findings were described in a father and a son with Nougaret-type CSNB, including undetectable rod response and decreased biphasic a-waves revealed by bright flash ERG, with a positive b/a-wave amplitude ratio and at least 50% greater than normal amplitudes [19]. Their LA 30-Hz flicker responses were normal. Notably, a characteristic feature was larger amplitudes of bright flash ERG than could be generated by cone responses alone. In addition, it was demonstrated from detailed ERG data that Nougaret-type CSNB is characterized by 100-fold reduction in rod sensitivity [19]. Our ERG findings (Fig. 3) of patients II-2 and II-3 were similar to those of Nougaret-type CSNB, but with minor differences. The a-waves from DA 3.0 ERG were delayed and broadened rather than biphasic, and b/a-wave amplitude ratio was negative. However, clear on and off responses were observed, which is inconsistent with Schubert–Bornschein-type CSNB or complete-type CSNB [20, 21]. Further, our pathognomonic ERG findings are different from Riggs-type CSNB, which is characterized by DA 3.0 and/or DA 10.0 ERG with minimal a-waves as well as b-waves derived only from cone responses. In other words, Riggs-type CSNB exhibits essentially no rod signal. Our comprehensive ERG results were therefore most compatible with those documented in Nougaret-type CSNB. The slight differences in the ERG findings might contribute to the difference in the flash intensity used between the previous [19] and current studies.

Previously, heterozygous missense GNAT1 variants have been reported in both Nougaret-type (p.G38D) and Riggs-type CSNB (p.I52N and p.Q200E) with autosomal dominant inheritance (Table 1) [2, 6,7,8, 19], while homozygous missense (p.D129G) and truncated GNAT1 variants (p.Q302* and p.C321*) are known to cause Riggs-type CSNB and progressive retinal degeneration with autosomal recessive inheritance, respectively (Table 1) [22,23,24]. So far, the GNAT1 variant (p.G38D) is the only documented cause of Nougaret-type CSNB. Experimental studies using transgenic mice with p.G38D have revealed two crucial deficiencies: poor ability to activate PDE6 (rod cGMP phosphodiesterase) and partially decreased GTPase activity in rod phototransduction [25, 26]. On the other hand, in Riggs-type CSNB, constitutive activation of rod phototransduction by impaired GTPase activity is suggested as the pathogenic mechanism leading to the absence of rod function [6]. In Nougaret-type CSNB, although it is unclear why the rod sensitivity and rod ERG responses are extremely reduced, the mutant (p.G38D) α-transducin in heterozygotes can be activated only by strong light stimuli, resulting in at least 50% greater than normal hyperpolarizing broadened a-waves in DA 3.0, DA 10.0, and DA 200 ERG (Fig. 3). These findings suggest that transducing activity in heterozygotes with the p.G38D variant might be suppressed incompletely.

Table 1 Clinical and genetic characterization of patients with reported GNAT1 variants
Full size table

Biallelic variations in the ABCA4 gene are responsible for ABCA4-associated retinopathies including STGD1, COD, and CORD [9,10,11,12]. Previously, phenotypic subtypes have been classified into three groups based on ERG abnormalities: normal full-field ERG in Group 1, normal rod ERG but reduced cone ERG in Group 2 (COD phenotype), and decreased both rod and cone ERG in Group 3 (CORD phenotype) [27]. In genotype–phenotype correlations of ABCA4-associated retinopathies, it has been reported that missense variants are associated with milder and later onset phenotypes, whereas null/truncated alleles are associated with more severe and earlier onset phenotypes [10]. Fujinami et al. [28] report that a higher proportion of ABCA4-associated retinopathy patients with childhood-onset phenotype were in Group 3 compared with adult-onset phenotype. Our childhood-onset CORD patients (III-1 and III-2) had biallelic truncated ABCA4 variants as well as the GNAT1 variant (p.G38D). In patient III-2, the cone and 30-Hz responses, which cannot be affected by the GNAT1 variant, were severely decreased (Fig. 3), indicating that the phenotype of ABCA4-associated retinopathy is classifiable into either Group 2 or Group 3. Although it is uncertain whether non-recordable rod responses in patients III-1 and III-2 were derived from either the GNAT1 or ABCA4 variants, the DA 200 ERG responses were much more decreased than those of patients (II-2 and II-3) with Nougaret-type CSNB alone (Fig. 3), suggesting that the biallelic ABCA4 variants themselves may contribute much more to rod dysfunction than the GNAT1 variant itself. Our ERG findings revealed that the biallelic ABCA4 variants alone are sufficient to develop the CORD phenotype seen in patients III-1 and III-2. This confirmed that patients III-1 and III-2 belonged to Group 3 [27]. Further, our FAF and OCT findings (Fig. 4) were consistent with ABCA4-associated retinopathy. Notably, the hyperreflective line corresponding to thickened ELM in OCT we observed in patient III-2 is one of the features previously observed to characterize childhood-onset ABCA4-associated retinopathy [16].

Patients suffering from simultaneous occurrence of STGD1 and other IRDs such as CSNB and X-linked ocular albinism (OA) have been reported previously. Huynh et al. [29] report that a female patient exhibiting a complex phenotype with both STGD1 and complete-type CSNB carried biallelic ABCA4 variants (p.L1201R and p.R2077G) and biallelic GRM6 gene variants (c.50_64del and c.1835_1837del). Lee et al. [30] report two female patients with STGD1 and carrier status of OA, who exhibited bull’s eye macular lesions, elevated FAF signals, and retinal mosaic patterns seen in OA, and carried biallelic ABCA4 variants (p.L541P and p.G1961E) and a heterozygous GPR143 gene variant (p.Y257C). Apart from IRDs, a large cohort study has recently reported that nearly 5% (101/2076) of patients with genetic disorders, who underwent WES, have two or more disease loci/disease-causing variants [31], indicating the importance of identifying concurrent phenotypes in a single pedigree. Our ERG findings (Fig. 3) of patients III-1 and III-2 with two monogenic disorders, Nougaret-type CSNB and childhood-onset CORD, were associated with coexistence of the GNAT1 and biallelic ABCA4 variants. This clearly reflects the principal that developmental mechanisms of complex phenotypes can be resolved only when multiple disease-causing gene variants are identified. On the other hand, we cannot exclude the possibility that the childhood-onset CORD phenotype could occur by interactions of the respective mutated proteins of ABCA4 and GNAT1, because both are co-localized to rod photoreceptors. This suggests that, where there are two different IRD phenotypes in a family, coexistence of two gene variants should be considered.

In conclusion, this is the second family of Nougaret-type CSNB with the GNAT1 variant (p.G38D). Also, our novel findings suggest that coexistence of the GNAT1 and biallelic ABCA4 variants is likely to be associated with an overlapping phenotype with both Nougaret-type CSNB and childhood-onset CORD.